😺 Google's Ultra AI power move

PLUS: Why you CAN'T trust AI 100%...

Welcome, humans.

Media’s always been a tricky business, but after years of private equity rollups, newspapers are struggling to stay afloat—and some are taking dangerous AI shortcuts to fill their pages.

Exhibit A: The Chicago Sun-Times just published a “Summer Reading List for 2025” featuring completely made up books supposedly written real authors.

As 404 Media discovered, Isabel Allende's “first” climate fiction novel “Tidewater Dreams”? Doesn't exist. Andy Weir's “The Last Algorithm”? Pure AI hallucination (but we’d totally read that if you wrote it, Andy!).

The plot thickened when it was revealed that Hearst subsidiary King Features created the insert, and nobody at Chicago Public Media (the Sun-Times' owner) even reviewed it before publication. They literally printed AI-generated content without a single human glance!

As a result, Chicago Public Media announced they'll now require “internal editorial oversight” for such content—essentially admitting their groundbreaking discovery that maybe newspapers should read things before publishing them. Revolutionary concept!

People: this is a textbook AI fail. When you ask AI to conduct a long task for you, without fact checking it, it is bound to introduce some hallucinations.

You know the old Reagan quote “trust, but verify?” When it comes to AI, that slogan should be “DON’T trust… verify.”

Here’s what you need to know about AI today:

Google launched everything at I/O.
Kurzweil's robot startup gets $100M investment.
Trump signs law criminalizing deepfakes.
Nvidia launches new AI tools to advance humanoid robot development.

Advertise in The Neuron here.

Google’s $250 Power Play & the Death of Search

Google just unleashed a torrent of AI announcements at I/O 2025, including a premium $250/month “Ultra” tier that reshapes the consumer AI landscape. Far from playing catch-up, Google is now solidly positioning itself as the AI leader—and is ready to monetize that advantage.

The headliner: AI Mode is now available to everyone in the US as a new search tab (and you can shop with it, too). Google made a bold statement that the future of search is AI-first, and they're not about to let competitors (like Perplexity, Grok, ChatGPT) eat their lunch.

Here’s the rest of the I/O highlights:

Google AI Ultra launches at $249.99/month (intro price: $124.99) with VIP access to all Google's AI features.
AI Premium rebranded to Google AI Pro ($19.99/month), keeping most existing features.
Gemini 2.5 Pro tops the LLMArena leaderboard in all categories, outperforming competitors.
Gemini 2.5 Pro Deep Think delivers impressive reasoning capabilities with parallel thinking techniques, scoring high on math and coding benchmarks.
Jules coding agent enters public beta, allowing anyone to try Google's asynchronous coding assistant for free.
Agent Mode is coming to Chrome, Search, and the Gemini app for task automation across websites.

Oh, and this isn’t AI related, but its a bit of a hobby horse for this remote-first team: Google Beam (formerly Project Starline) is becoming a real product through HP later this year—this is gonna be sick!!

As you can see, the sheer volume of announcements shows Google is executing on a comprehensive strategy that spans consumer subscriptions, developer tools, and ambitious research.

Here’s what this means for you: The Ultra tier's steep pricing shows Google believes power users will pay significant premiums for cutting-edge capabilities—creating a new “luxury tier” in consumer AI.

The global AI Mode rollout in search fundamentally changes how we retrieve information. Users are already asking 2-3x longer questions, with Google promising this will drive even more web traffic rather than cannibalizing it—a claim that remains to be validated (ahh, validation… the theme of today’s newsletter, apparently!!).

And Google's technical advancements are equally impressive. Gemini 2.5 Pro Deep Think uses parallel reasoning techniques to consider multiple questions, scoring impressively on the 2025 USA Mathematical Olympiad and achieving 84% on the MMMU benchmark. Translation? Your AI will be able to handle ever more complicated queries.

Our take: The battle lines in consumer AI are now clearly drawn. This is a war between Google and OpenAI. The two combine for an amount of AI market share already that even the likes of Microsoft and Meta would struggle to topple. It also leads us to wonder how much longer Anthropic and other labs can remain competitive (our completely uninformed bet is Anthropic eventually gets acquired by Amazon… is there a polymarket poll for this yet? Cause there should be!).

Google is leveraging its scale to accelerate AI adoption. It’s also betting consumers will pay significantly more for AI that delivers tangible results. One key difference between Gemini Ultra and ChatGPT Pro ($200/mo) is the latter was marketed more to power users in research applications and similar fields. Google, on the other hand, seems pretty confident AI enthusiasts will be comfortable with that number.

Here’s the full video of the I/O keynote in case you have three hours and no plans.

Google I/O '25 Keynote

For everyone else, watch this if you only have 32 mins or this if you only have 2 mins.

FROM OUR PARTNERS

CodeRabbit’s Free AI Code Reviews in your IDE—VS Code, Cursor, Windsurf

CodeRabbit is the AI code review platform purpose built to speed up code reviews and improve code quality. Integrating seamlessly into git workflows and existing toolchains, it delivers codebase-aware reviews and supports all programming languages.

Now they are excited to deliver AI code reviews directly in VS Code, Cursor, and Windsurf–letting you and your team (vibe) code with confidence and review while keeping in flow.

Even better: Code reviews in the IDE are free! That’s right: line-by-line reviews and one-click fixes, all in your IDE at no cost to you.

Install the extension and start vibe checking your code today!

Prompt Tip of the Day

After that Chicago Sun story, this advice is more important than ever: When you’re really convinced you’re right… that’s the best time to double-check your logic. ChatGPT can help you stress test your thinking before someone else does.

Try this:

“Here’s what I believe: [insert opinion]. What’s the strongest argument against it (slash help me “steel man” the other side)? How would a smart, skeptical person respond?”

It’s like instant access to a thoughtful contrarian. You’ll either strengthen your position—or realize you’ve been building your case on vibes and caffeine.

Treats To Try.

Gemma 3n from Google runs powerful AI directly on your phone or tablet with faster response times, allowing you to process text, images, audio, and video offline while preserving your privacy—try it here in AI Studio (read more).
GitHub Copilot’s new coding agent turns your issues into completed pull requests while you focus on high-value work.
NotebookLM’s new mobile app brings AI-powered information understanding to your phone, anywhere.
eSelf AI creates lifelike video agents that understand and interact with customers in 30+ languages.
Overlap automates your video marketing workflow by turning long-form content into engaging clips.
Clemta Intelligence transforms your business finances by automating bookkeeping, tax preparation, and financial reporting.
Mailgo discovers ideal leads and launches high-deliverability cold email campaigns.
Replicate lets you run, fine-tune, and deploy AI models across different domains with just one line of code through a simple, powerful API.

See our top 51 AI Tools for Business here!

*This is sponsored content. Advertise in The Neuron here.

Around the Horn.

Nvidia CEO Jensen Huang advised students to learn AI prompting skills for career success.
Ray Kurzweil's humanoid robot startup Beyond Imagination secures $100 million investment.
Trump signed the Take It Down Act, criminalizing nonconsensual intimate images and AI deepfakes. The law requires social media platforms to remove such content within 48 hours.
Nvidia introduced Isaac GR00T N1.5, a foundation model for robot reasoning, and GR00T-Dreams, a blueprint for generating synthetic motion data.
Senator Grassley introduced the AI Whistleblower Protection Act to shield AI employees who report industry risks.
Google's Jeff Dean predicted AI could match junior coders' skills within a year.
NetNation acquired Yola, an AI-powered website builder with over 15 million users.

Midweek Wisdom

MIT expert Daniela Rus highlighted AI's transformative potential across sectors like manufacturing, agriculture, and healthcare.
Venture capitalist Eze Vidra outlined how AI has transformed the startup landscape, making product development easier but intensifying competition and raising the bar for differentiation.
Harvard's Galileo Project is using advanced AI technology to search for extraterrestrial evidence—this is attracting academic research and Pentagon interest in what was once considered a fringe scientific pursuit.
A newly-published study reveals that most large language models tend to overgeneralize scientific findings—but not erroneously (also, newer models may be more likely to broaden claims beyond the original research scope).
Stephen Wolfram (math / science genius) explored the inner workings of ChatGPT, revealing that language might be fundamentally simpler than previously thought since AI implicitly discovered underlying patterns in human communication.
One final reminder to verify any and all Ai outputs—Check out this physicist's candid account of how AI-for-science research suffered from widespread overoptimism—he found 79% of papers claiming AI superiority used weak baselines, while negative results almost never got published, creating a distorted picture of AI's actual capabilities in scientific applications.